The statistical tools that have been presented in the preceding chapters should be used together to improve quality. Why is that? Assume that a process is out of control but an experiment is performed to compare the standard operating process with an experimental process that is conjectured will produce superior results. If the standard process is used for, say, one month and the experimental process is used for the same length of time, the results will not be comparable if the process is out of control in such a way as to affect the results of the experiment.
For example, if one or more process characteristics are out of control and this condition causes process yield to decrease when the experimental process is being used, a process engineer might be misled into perhaps falsely concluding that the experimental process is not superior. Similarly, we could erroneously conclude that a process change has been beneficial when we are actually seeing a data shift that is due to a parameter change. This was illustrated in Section 13.2 and is discussed further in Section 17.1.
Broadly stated, there is a need for a climate of “statistical thinking” to exist in order for combinations of statistical techniques to be used effectively. Britz, Emerling, Hare, Hoerl, and Shade (1997) gave the following definition of the term:
Statistical thinking is a philosophy of learning and action based on the following principles: